Entity Disambiguation research benchmarks

Benchmarks

Dataset NameSOTA methodMetricTrendResultsLast Updated

MSNBC (MSB) (test)

CHOLAN

83.4Micro F1

Feb 26, 2026

CoNLL table P (test)

OURS

96.7Accuracy

Feb 26, 2026

tweet set

DSRM1234

71.89Micro Precision@1.0

Feb 26, 2026

MHERCL-ITA (test)

DELICATE_GBT+LLM

70.44Accuracy (Micro)

May 25, 2026

ENEIDE AMD 1.0 (test)

DELICATE_GBT+LLM

72.67Micro Accuracy

May 25, 2026

ENEIDE DZ 1.0 (test)

DELICATE_GBT+LLM

63.05Accuracy (Micro)

May 25, 2026

6 standard ED

KBED

90Avg ED F1

Feb 26, 2026

CoNLL alias table H (test)

OURS

92.5Accuracy

Feb 26, 2026

AIDA-B

REL (2014)

85.5Macro F1

Feb 26, 2026

AIDA-B (test)

CHOLAN

85.7Micro F1

Feb 26, 2026

GERBIL (test)

De Cao et al.

88.8F1 Score

Feb 26, 2026

ISTEX

XLNet + 1-hop KG Context

0.964Precision

Feb 26, 2026

BLINK (controlled mentions)

Module-Based Approach

91.6Accuracy (< 5 Candidates)

Apr 6, 2026

WikiLinksNED Unseen Mentions (test)

Cao et al. (2020)

64.3Seen Accuracy

Feb 26, 2026